Differentiating motivation and control in the Theory of Planned Behavior

Ajzen has distinguished between self-efficacy (i.e., confidence, ease-difficulty) and controllability (i.e., control, up to me) item clusters when measuring perceived behavioural control (PBC). Rather than a two-factor control construct, we have suggested that confidence and ease-difficulty items possess measurement complexity between intention and PBC and do not measure Ajzen's PBC construct as cleanly as do control items unless a phrase is included to hold motivation constant (e.g., if I was really motivated…). The purpose of this study was to include PBC items with and without phrases that hold motivation constant and test the effects of measurement redundancy on other TPB coefficients when predicting/explaining exercise intentions. Participants were 241 undergraduate students who completed measures of the TPB. Results supported our theorizing for a single control construct by showing that confidence and ease-difficulty items without motivation held constant displayed factor complexity between PBC and intention, but loaded exclusively on PBC when motivation was held constant. Conversely, control items loaded exclusively on PBC regardless of whether motivation was held constant. In the TPB model where the factor complex items were used to represent PBC, PBC had a very large (β = 0.73, p < 0.01) effect on intention while affective attitude had a small effect (β = 0.12, p < 0.05). In the TPB model where factor clean items were used to represent PBC, PBC had a much smaller effect (β = 0.38, p < 0.01) on intention, affective attitude had a much larger effect (β = 0.42, p < 0.01), and the effect of subjective norm became significant (β = 0.15, p < 0.05). We discussed the results in terms of how PBC measurement with Ajzen's denoted self-efficacy items may bias basic and applied TPB research unless motivation is held constant.